标题:Recommendation Based on Frequent N-adic Concepts
作者:Wang, Di; Ma, Jun
作者机构:[Wang, Di; Ma, Jun] Shandong Univ, Sch Comp Sci & Technol, Jinan 250101, Peoples R China.
会议名称:16th Asia-Pacific Web Conference (APWeb)
会议日期:SEP 05-07, 2014
来源:WEB TECHNOLOGIES AND APPLICATIONS, APWEB 2014
出版年:2014
卷:8709
页码:318-330
DOI:10.1007/978-3-319-11116-2_28
关键词:Recommendation; Frequent N-adic Concepts; Tags; Contexts
摘要:In social networks, many users tend to share items such as movies, books, songs and images by rating them with a series of discrete numbers or annotating them with a set of tags. Clearly, there are some semantic relationships among the users, items, ratings, tags and other information. Most of the past works only focused on some ternary relationships such as users-items-ratings or users-items-tags to make recommendations. But the ternary relationships which do not make good use of the given information are insufficient to provide accurate recommendations. In this paper, we propose a novel recommendation method based on frequent n-adic concepts which can mine the hidden conceptualization in the relationships. If there are tags, we model the relationships into the quadruples and if there are no tags, we also have some other information and model the relationships into the quintuples . Experimental results on MovieLens dataset demonstrate that our method has shown a significant improvement over the state-of-the-art recommendation approaches in terms of precision.
收录类别:CPCI-S;EI;SCOPUS
资源类型:会议论文;期刊论文
原文链接:https://www.scopus.com/inward/record.uri?eid=2-s2.0-84958536622&doi=10.1007%2f978-3-319-11116-2_28&partnerID=40&md5=2a11b3dba2e1e542b3598b656d973018
TOP